[Elecraft] start without ANY extra roofing filters....

[Jack Smith]

Hat tip to Larry ... now all I need is a commission on every QEX  sold ...

Crystals are non-linear and their motional parameters are, to some 
degree or other, a function of drive voltage. Since a filter's loss is a 
function of its motional parameters, the corollary to that is 
intermodulation can and will be caused by what we think of as purely 
passive elements such as crystals. (This phenomenon is also found in 
ferrite cores and powdered iron core inductors, as they have a 
non-linear B versus H curve.)

There are examples of receivers that have as the limiting IP3 crystal 
filter intermodulation. See Experimental Methods in RF Design for a 
discussion of Wes Hayward's observation of crystal filter IMD when 
building a receiver featured in that book.  It's devilishly hard to 
measure crystal filter IMD, however, for a variety of reasons.

This is why a filter with fewer elements (poles) can, in some 
circumstances, yield a better IP3 than a filter with more poles, as 
counter-intuitive as that might seem. Whilst the filter with more poles 
will keep more trash out of later receiver stages, small changes to the 
motional parameters of the  crystals that make up the filter with more 
poles will have a greater effect on the filter's transfer function than 
for a filter with the same crystals but fewer poles. Thus, although 
later stages are better protected from undesired signals, that very 
protection itself causes intermodulation interference.

That's why a high performance receiver must be designed in a holistic 
fashion.

Jack K8ZOA



Larry Phipps wrote:
> Well done, Bill. This again shows the importance of real life 
> measurements... and why I didn't order any filters until more is known 
> about them.
>
> Anybody interested in the subject of filter design must read the 
> article by list member Jack, K8ZOA in the current QEX. It gives a lot 
> of valuable xtal filter design insight, and has a page of excellent 
> references at the end for those who wish to read more on the subject. 
> This is a complex subject, but as Jack points out, proper 
> characterization of the crystals and rigorous attention to detail can 
> produce accurate models and repeatable designs. Jack touched on drive 
> level dependency in his article. Perhaps he can focus in on the 
> effects of xtal nonlinearity as it affects IMD for a future piece (not 
> trying to create work for you Jack ;-)
>
> This is a subject which seems to be gaining in importance as receiver 
> designs surrounding the xtal filter seem to be improving to the point 
> where the filters are becoming the limiting factor in IMD performance.
>
> 73,
> Larry N8LP
>
>
>
> Bill Tippett wrote:
>>
>>
>> I wrote:
>>
>> >          Bottom line:
>>
>> 1.  Narrower is not always better (Ten-Tec experience)
>> 2.  8-poles is not always better than 5-poles (per Inrad)
>> 3.  Let IMD and BDR measurements be your guide
>>
>>         More evidence below to support waiting for IMD/BDR 
>> measurements before ordering any roofing filters.
>>
>>                                 73,  Bill  W4ZV
>>
>> http://mailman.qth.net/pipermail/1000mp/2007-April/020755.html
>>
>> There is a fascinating article describing IMD tests on the IC-7800 by 
>> DC4KU
>> in CQ-DL, August 2005 (in German). In these tests, IP3 at 2 kHz offset
>> degrades by an astounding 16 dB when switching from the 15 kHz to the 
>> 6 kHz
>> roofing filter. This degradation is due to passive IMD in the filter, 
>> and
>> possibly also to IMD in the filter driver amplifiers caused by 
>> mismatch when
>> the filter is excited outside its passband. I can send you an
>> English-language summary of the relevant part privately, if you wish.
>>
>> It is highly significant that professional receivers manufactured by the
>> likes of R&S, Rockwell-Collins, Racal and Harris have a single roofing
>> filter. This filter is typically 12 to 16 kHz wide, to pass 
>> multi-channel
>> ISB, VFT (multiplexed teletype) and high-speed crypto, all of which have
>> extremely stringent in-band IMD requirements. To quote a British 
>> engineer
>> who used to design shipboard HF receivers for the Royal Navy:
>>
>> The up-converting architecture, with a roofing filter at a first IF 
>> above
>> the highest RF frequency, allows the designer to limit the bandwidth
>> presented to the first IF chain and second mixer. The bandwidth of this
>> filter is a trade-off. Its 3 dB BW must be sufficient to pass the widest
>> emission the receiver is required to handle, but not so narrow that 
>> IMD and
>> temperature-drift effects in the filter become a concern.
>>
>> Cheers for now, 73,
>> Adam VA7OJ/AB4OJ
>>
>>
>>
>> _______________________________________________
>> Elecraft mailing list
>> Post to: Elecraft at mailman.qth.net
>> You must be a subscriber to post to the list.
>> Subscriber Info (Addr. Change, sub, unsub etc.):
>> http://mailman.qth.net/mailman/listinfo/elecraft   Help: 
>> http://mailman.qth.net/subscribers.htm
>> Elecraft web page: http://www.elecraft.com
>>
>>
> _______________________________________________
> Elecraft mailing list
> Post to: Elecraft at mailman.qth.net
> You must be a subscriber to post to the list.
> Subscriber Info (Addr. Change, sub, unsub etc.):
> http://mailman.qth.net/mailman/listinfo/elecraft   
> Help: http://mailman.qth.net/subscribers.htm
> Elecraft web page: http://www.elecraft.com
>